274 ALASKAN GLACIER STUDIES
stones and tree trunks, the latter, some of which were over two feet in diameter, being inverted, prostrate, inclined, or beginning to be overturned. Against this moraine the ice cliff of the glacier extended closely in 1909 (PI. CX, A), being at least as far out as the advance that killed the trees in 1892, and perhaps a little farther. The glacier terminated against this push moraine, above which it towered precipitously, but it was not so severely crevassed that one could not walk out a short distance upon its surface. At one point we could enter some of its larger crevasses and see the deep blue of the ice from within. The dirty basal ice projected farthest in this area, there being no projecting splinters of clear upper ice thrust forward among these trees, as there were in the western timber belt.
Outside the push moraine, in the tangle of tree trunks, there were a few rolls of peaty soil with shrubs and other plants, still growing in 1909, although often at right angles to their normal vertical position. One such roll was a typical anticline overturned on the side away from the thrust, whose effect here was felt at least 100 feet from the ice. Few of the growing trees which were not affected by the 1892 advance had been overturned in August, 1909.
At the eastern end of this forest belt a small pool was held in between the ice front and the rocky slope of the higher part of the island (P\. CX, B), and its outlet flowed westward through the edge of the forest to the small pond near the heath mentioned above. Out of this east pool rose a series of overturned and compressed folds of peaty soil on the extreme eastern edge of the belt of overturned trees. On the opposite (east) side of this pool the glacier had pushed up a former sea bottom deposit of clay with abundant marine shells.
In 1910 the whole situation was changed, owing to the advance of the glacier. A few of the inclined trees were recognizable, but the ice front had advanced several hundred feet into the growing forest and had heaped up among the growing trees a terminal moraine mass of tree trunks, soil, rock, and vegetation (PI. CVI, B). It rose to a height of 30 or 40 feet, at least, where ten months before there had been no surface accumulation. Above this heap rose the inclined trees in a confused thicket, with trunks pointing at every conceivable angle. The growing trees in this eastern timber belt had not been overridden and buried, as in the western timber belt, nor pushed forward without being overturned, as was the case with the scattered trees on the intermediate push moraine back of the heathy meadow. All of them were overwhelmed by the advancing glacier, but only a few were overridden, while most were heaped up in the windrow of forest debris, whose edge advanced steadily into the growing forest from July to September, 1910. Some distinctive blackened tree trunks, which had been killed previously, changed position from August, 1909, to July, 1910, not only by forward movement but by being lifted vertically 80 or 40 feet.
Along the eastern edge of this forest belt, in 1910, the site of the little pool observed in 1909 was overridden; and the glacier had advanced to the edge of a high bank which was a hundred feet or more from the ice the year before. As yet there were no push moraine or tangle of tree trunks at the margin of the glacier, which had advanced across a clear space formerly occupied by the eastern pool and by a marginal stream course. In July and September, 1910, however, the forest was just being invaded by the ice. One tree which was seven feet from the ice on July 6th was touched by the glacier on July 9th, the advance being 2$ feet a day; and before September 5th this tree was over-